Debian web pages - what needs to be done
Here is a list of things that are on the TODO list for the web pages.
Items that need some programming have a p at the beginning and those
that are ongoing commitments have a c. Don't be put off by the length.
Most of these are one shot deals.
Feel free to volunteer for any of them.
c Better info about releases. For example, what kernel is shipped with
2.0. We need to find out what information people most want and make
sure that it is easy to find.
p people.html is too large and needs to be split into a few pages
This should probably wait until the developer db is up and running
as the scripts will need to be rewritten then anyway.
The navigation bar at the top of the pages is language specific.
We either need to set up an easy method to create translated images
or to get rid of it.
c security pages are way out of date. This really should be taken care
of by the security team but I don't know what is going on with them.
There should be pages specific to each port of Debian. There are
some available which can be linked in for the short term. Ideally
a framework will be created so that there is a similar look to each
ports pages. Once the framework is done, a person involved with each
port should take care of their section.
The web pages need to use push mirroring. I believe that Jason Gunthorpe
is already organizing this.
c Someone to coordinate all the translations. We need to make it as easy
as possible for people to translate. They would communicate with the
organizers of translations to solve any problems and grant write access
to CVS. This person should also actively seek translators for new languages.
A few other people should recieve mail to webmaster. Maybe it should
simply be forwarded to debian-www.
Most of the docs in the Debian Documentation Project (DDP) are on an
ex-developers machine. These need to be moved to www.debian.org and
put under CVS. This is really the domain of the doc group, but as these
documents are also available under the web, I've added it here.
p The news page needs to be rewritten so that each piece of news is in it's
own file (under directories by year). This will allow translations of
individual news items, like announcements of new releases, without requiring
all items to be translated. It would probably save time in the long run
if the index was automatically generated from the files.
The vendors page needs to be have the entries organized by country with links
at the top into each countries listing. Easy.
CVS should be set up so that changes to the english version of a file cause
mail to be sent to debian-www with the name of the file changed and the message
log. This would allow translators to decide if the translated page needs to be
updated immediately or to wait when the changes are minor.
The introductory pages to Debian could be better written.
analog, which is used to generate reports from apache web logs, needs to use
cache files to speed up the generation of reports. I'll probably do this
myself.
The Packages section needs to handle multiple ports.
p Links into the archive should allow the selection of the site to download from.
This can be done rather slickly using wml by generating one version of the form
and simply including it into a file at the right place. Even the form will be
generated automatically once the new version of the mirror list is far enough
along. I'll probably do this one myself.
p Guy currently handles the mail list archives. Since he is so busy, it would
save him time and allow changes to be made more quickly if someone else takes
care of this. This requires a good understanding of perl and mail.
p Searching on the web site needs to be done under one program. Different areas
may have slightly different needs and search customizations, but should allow
a global search of everything. Everything is currently a mishmash: Packages section
uses htdig (doesn't do sub-string searches well), the mailing lists use glimpse
(a pain to administer and has its own limitations), and the only way to do a
global search on the bug lists is by using the searching linked on the main
web page which uses a third party.
c A URL checker needs to be run on the site once a week or so. This recursively goes
through the site checking for broken links. Doing this is one of the best ways
to avoid complaints.
Jay Treacy
Reply to: